A method and system for village-in-city reconstruction planning based on a generative model

Abstract

The application provides a village-in-city reconstruction planning method and system based on a generative model, and the method comprises the following steps: according to determined urban conditions, village-in-city conditions and a reconstruction scheme set, a first generative model is used to generate urban activity conditions after m years; according to the urban activity conditions after m years, a second generative model is used to calculate urban internal commuting conditions after m years and urban facility utilization conditions after m years; economic income and expenditure conditions after m years are obtained according to the reconstruction scheme set, and a heuristic algorithm is used to perform multiple iterations according to the economic income and expenditure conditions after m years, the urban internal commuting conditions after m years and the urban facility utilization conditions after m years, so as to obtain multiple adjusted target reconstruction schemes; and the multiple adjusted target reconstruction schemes and corresponding total economic income and expenditure conditions within m years, total urban facility utilization conditions within m years and total urban internal commuting conditions within m years are analyzed to obtain an optimal target reconstruction scheme.

Description

Technical Field

[0001] This invention relates to the field of urban planning technology, and in particular to a planning method and system for the redevelopment of urban villages based on generative models. Background Technology

[0002] Urban village redevelopment refers to the comprehensive transformation of urban villages based on regional socio-economic development and overall urban planning, in accordance with urbanization requirements. Actively and steadily promoting urban village redevelopment in megacities and super-large cities is an important measure to improve people's livelihoods and drive high-quality urban development.

[0003] However, large-scale urban village renovation in megacities and super-large cities often requires overall planning across hundreds or even thousands of urban villages, determining the order and intensity of renovation, coordinating fund allocation within the city, considering the financial balance, facility layout, and housing price changes after the implementation of the renovation plan, and taking into account the impact of specific renovation plans on the city's overall population distribution, job-housing balance, and transportation demand. This poses a huge challenge to traditional planning decision-making methods.

[0004] Traditional methods of urban renewal planning cannot extrapolate changes in urban factors such as housing prices, infrastructure, and population, nor can they simultaneously optimize and compare multiple renewal objectives, thus failing to meet the needs of planning decisions. Summary of the Invention

[0005] This invention provides a planning method and system for urban village redevelopment based on a generative model, which addresses the shortcomings of existing urban village redevelopment planning methods that cannot simultaneously optimize and compare multiple redevelopment targets, thus failing to meet the needs of planning decisions.

[0006] This invention provides a planning method for urban village redevelopment based on a generative model, comprising:

[0007] Based on the established urban conditions g, urban village conditions u, and redevelopment plan set P, the pre-trained first generative model is used to generate the urban activity status g m years later. m The first generative model is trained based on urban condition training samples, urban village condition samples, and renovation scheme sample sets.

[0008] Based on the urban activity situation m years later g m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I mThe second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m ;

[0009] Based on the set of renovation plans P, the economic income and expenditure situation E after m years is obtained. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Multiple adjustments to the target modification scheme were obtained by using heuristic algorithms for multiple iterations.

[0010] Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years, the optimal target renovation scheme is obtained.

[0011] According to the urban village redevelopment planning method based on generative models provided by the present invention, the first generative model includes a first network and a second network;

[0012] Based on the established urban conditions g, urban village conditions u, and redevelopment plan set P, the pre-trained first generative model is used to generate the urban activity status g m years later. m Specifically, it includes:

[0013] Given the established city condition g, noise is extracted from a standard Gaussian distribution and then added to the established city condition g to obtain the current random city condition. ;

[0014] Random city conditions Compared with the predefined maximum number of denoising steps The input is fed into a pre-trained first network for processing to obtain the predicted noise. ;

[0015] Random city conditions The conditions of the urban village (u) and the set of renovation plans (P) are input into a pre-trained second network to calculate the gradient of the loss function. ;

[0016] Update the noise based on the gradient value. The current random city status is updated based on the updated noise to obtain the updated random city status.

[0017] The process iteratively executes the steps of inputting the updated random city conditions into the first network and the second network respectively to obtain the next updated random city conditions, until the maximum number of denoising steps T is 0, thus obtaining the city activity status after m years.

[0018] According to the urban village redevelopment planning method based on generative models provided by the present invention, the training method of the first network includes:

[0019] Randomly select a city condition training sample from the training sample dataset;

[0020] Determine the random sampling time. and from a constant sequence Extract the corresponding constant from ;

[0021] A random sample noise is randomly drawn from a standard Gaussian distribution of the same size as the city condition training sample, and the random sample noise and the constant are added to the city condition training sample to obtain the noisy city condition training sample.

[0022] The noisy urban condition training samples and randomly selected time The noise is input into the first network to obtain the predicted noise;

[0023] The loss value is calculated based on the predicted noise and random sample noise. The parameters in the first network are updated using the predicted loss value through stochastic gradient descent until the parameters in the first network converge.

[0024] According to the urban village redevelopment planning method based on generative models provided by the present invention, the training method of the second network includes:

[0025] Randomly select a city condition training sample from the training sample dataset. ;

[0026] Training samples of randomly selected urban conditions Starting from the corresponding time point, proceed backwards to the time point at the beginning of the sample dataset as the ending point, and randomly select a time interval. ;

[0027] Based on the sampling time interval Obtain known samples of urban village conditions and the corresponding set of renovation scheme samples ;

[0028] Randomly selected urban condition training samples The predicted modification time interval is obtained by inputting the data into the second network. Urban Village Situation and renovation plan ;

[0029] Calculate the predicted renovation time interval The time interval between sampling and the predicted conditions of urban villages Compared with known samples of urban village conditions and predicted redevelopment plans The difference between the sample and the known modification schemes is used to obtain the second loss value;

[0030] The parameters of the second network are updated using the second loss value through stochastic gradient descent until the parameters of the second network converge.

[0031] According to the urban village redevelopment planning method based on generative models provided by this invention, based on the urban activity status g after m years... m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m Specifically, it includes:

[0032] Define multiple agents in the city. For any agent in the city... It records the set of all the locations it has visited;

[0033] For any time Random numbers are drawn from a random distribution. ;

[0034] Calculate the exploration probability of agent i ,in, and These are two pre-determined hyperparameters. This represents the number of grid points that agent i has already visited;

[0035] If p≥ The control agent performs a return operation, randomly selecting a location from the set of locations it has visited and updating its location; the probability of visiting each location is proportional to the frequency with which it has been visited before.

[0036] like The control agent i performs an exploration operation, randomly selecting a location from the set of other locations in the city and updating its location; wherein, the probability of visiting each location is directly proportional to the number of people in that location, and the probability of visiting each location is inversely proportional to the distance between two locations;

[0037] Perform the above steps on all agents in the city until time. Reaching the specified time step ;

[0038] Count the number of visits to city facilities of interest, and calculate the utilization of city facilities after m years.

[0039] Count the number of passages between grids within the city and calculate the city's commuting situation after m years. .

[0040] According to the urban village redevelopment planning method based on generative models provided by this invention, based on the economic income and expenditure situation E after m years... m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m By using heuristic algorithms for multiple iterations, several adjusted target modification schemes were obtained, including:

[0041] Based on the economic income and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Calculate the dominance relationships between objective functions to obtain non-dominated and dominated modification schemes;

[0042] For the dominated modification schemes, the crowding is sorted according to the distance between the dominated modification schemes within each level, based on the degree of domination.

[0043] Based on the degree of domination and crowding of the dominated transformation plan, select The optimal modification scheme is used as the parent solution for the next generation;

[0044] Calculate the crossover probability between parent lines Mutation probability The same result can be obtained by exchanging or updating parts of the sequence of the modification scheme. The modification plan for each offspring is used as the solution for the next generation of offspring;

[0045] Using the parent solution and the child solution of the next generation as the updated modification scheme, and processing them through the first generative model and the second generative model, we obtain the updated economic balance E after m years. m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Continue iterating until the required number of iterations is met or the set of modification schemes P is no longer updated.

[0046] Output the results of the last iteration, resulting in multiple adjusted non-dominated target modification schemes.

[0047] According to the urban village redevelopment planning method based on a generative model provided by this invention, the optimal target redevelopment plan is obtained by analyzing the multiple adjusted target redevelopment schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years. Specifically, the optimal target redevelopment scheme includes:

[0048] For each adjusted target renovation plan, let g represent the corresponding urban activity status m years later. m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Statistical processing was performed separately to obtain the total economic income and expenditure E, the total urban facility utilization I, and the total urban commuting C within the m years.

[0049] The optimal target renovation plan is determined by calculating the total economic revenue and expenditure (E) over m years, the total urban infrastructure utilization (I) over m years, and the total urban commuting situation (C) over m years, along with their respective weights.

[0050] This invention also provides a planning system for urban village redevelopment based on a generative model, comprising:

[0051] The first parameter generation module is used to generate the urban activity status *g* m years later, based on the determined urban status *g*, urban village status *u*, and redevelopment plan set *P*, using a pre-trained first generative model. m The first generative model is trained based on urban condition training samples, urban village condition samples, and renovation scheme sample sets.

[0052] The second parameter generation module is used to generate the city activity status g m years later. m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m ;

[0053] The iterative calculation module is used to obtain the economic income and expenditure situation E after m years based on the set of transformation schemes P. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years Im Multiple adjustments to the target modification scheme were obtained by using heuristic algorithms for multiple iterations.

[0054] The target renovation plan generation module is used to analyze the multiple adjusted target renovation plans and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years to obtain the optimal target renovation plan.

[0055] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the urban village redevelopment planning method based on generative models as described above.

[0056] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the urban village redevelopment planning method based on a generative model as described above.

[0057] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the urban village redevelopment planning method based on generative models as described above.

[0058] The present invention provides a method and system for urban village redevelopment planning based on a generative model. Based on the determined urban conditions g, urban village conditions u, and a set of redevelopment schemes P, it uses a pre-trained first generative model to generate the urban activity status gm m years later. Then, based on the urban activity status gm m years later... m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m Then, using the economic balance and expenditure situation E after m years m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m This method utilizes heuristic algorithms for multiple iterations to obtain several adjusted target renovation schemes, and then selects the optimal one. By employing a generative model, this embodiment can quickly simulate and predict the future economic, social, and environmental impacts of different renovation schemes. This allows for the simultaneous consideration of multiple objectives, such as economic benefits, utilization of public service facilities, and urban commuting. Through multi-objective optimization, it finds the optimal balance point, providing decision-makers with more comprehensive and accurate data support, thereby enabling more scientific and rational planning decisions. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0060] Figure 1 This is one of the flowcharts of the urban village renovation planning method based on generative models provided by the present invention.

[0061] Figure 2 This is a schematic diagram of the planning framework for the urban village renovation method provided by the present invention.

[0062] Figure 3 This is a schematic diagram of the training method for the first network provided by the present invention.

[0063] Figure 4 This is a schematic diagram of the framework of the first generative model provided by the present invention.

[0064] Figure 5 This is a schematic diagram of the training method for the second network provided by the present invention.

[0065] Figure 6 This is the second flowchart of the urban village renovation planning method based on generative models provided by this invention.

[0066] Figure 7 This is a schematic diagram of the framework of the second generative model provided by the present invention.

[0067] Figure 8 This is the third flowchart of the urban village renovation planning method based on generative models provided by this invention.

[0068] Figure 9 This is a schematic diagram of the framework of the heuristic algorithm H provided by the present invention.

[0069] Figure 10 This is the fourth flowchart of the urban village renovation planning method based on generative models provided by this invention.

[0070] Figure 11 This is a schematic diagram of the urban village renovation planning system based on a generative model provided by the present invention.

[0071] Figure 12 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0072] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0073] First, the terminology used in this embodiment will be explained illustratively.

[0074] E (Economic Income and Expenditure): This indicator reflects the economic benefits of the renovation project. It may include profits from commercial housing development and rental income from rental housing. Economic returns are a crucial indicator for assessing the economic feasibility of a renovation project, as they relate to the project's return on investment and economic efficiency.

[0075] I (Urban Facility Utilization, or Moran Index): The Moran Index is a spatial autocorrelation statistic used to measure the degree of clustering of spatial data. Here, I refers to the utilization of renovated public service facilities, assessed using the Moran Index. A higher Moran Index indicates that the use of public service facilities exhibits strong spatial clustering, which may mean that the distribution and utilization of facilities are more efficient.

[0076] C (Urban Commuting): This indicator reflects the commuting situation within the city after the renovation, and may include commuting time, commuting distance, traffic flow, etc. Urban commuting is an important indicator for evaluating urban operational efficiency and residents' quality of life, and it involves the planning and optimization of the transportation system.

[0077] Generative models are a class of machine learning models capable of generating new data instances. They are typically based on probabilistic generation methods, learning the underlying distribution of data from training data and generating new data samples accordingly. In urban village redevelopment planning, the application of generative models includes: generating future urban activity at a specific point in time based on existing urban conditions, urban village conditions, and redevelopment plans. This includes economic activity, population distribution, and traffic flow; the model can then predict urban commuting and access to urban facilities after the redevelopment plan is implemented, helping planners assess the potential impact of different redevelopment plans; by simulating different redevelopment plans, generative models can help planners understand the impact of these plans on the future development of the city, thereby optimizing the plans.

[0078] Heuristic algorithms are a class of algorithms used to solve optimization and search problems. They are typically based on experience and intuition to guide the search process and find approximate solutions. Heuristic algorithms do not guarantee finding the global optimum, but in practical applications, they often find sufficiently good solutions within a reasonable timeframe. Heuristic algorithms are used to handle multiple adjusted modification schemes, optimizing them through multiple iterations to ultimately obtain the optimal target modification scheme. This involves calculating the dominance relationships between objective functions, crowding ranking, selecting the optimal solution as the parent, and generating new child solutions through crossover and mutation operations.

[0079] Non-dominated solution set: The non-dominated solution set contains solutions that are optimal on all objectives, representing trade-offs between different objectives. In the non-dominated solution set, no solution is strictly superior to all other solutions on all objectives. For example, consider an optimization problem with two objectives: minimizing cost and minimizing time. Suppose we have three solutions A, B, and C, with the following objective function values: A: cost = 10, time = 5; B: cost = 8, time = 6; C: cost = 9, time = 4. In this example: B is dominated by A and C because A has better cost and time than B, while C has better time than B. A and C do not dominate each other because they have trade-offs between objectives: A has better cost than C, but C has better time than A. Therefore, the non-dominated solution set is {A, C}.

[0080] The following is combined with Figures 1-10 This invention describes a generative model-based urban village redevelopment planning method and system according to embodiments of the present invention.

[0081] In this embodiment, for a given urban area, it is divided into A fixed-range spatial grid, Each grid has a corresponding set of attributes. These correspond to the size of the resident population, the price of second-hand housing transactions, the price of rental housing, and the number of facilities organized by type within the urban grid.

[0082] Based on this, the areas in the city that need to be renovated are identified. The spatial extent of an urban village is represented as... For each urban village, it possesses a set of core attributes. This includes the corresponding grid coordinates, land area, building volume ratio, and the number of facilities organized by type.

[0083] The planning and formulation task is within the year range to be planned. Inside, for Each urban village to be planned will have a corresponding development plan, forming a complete set of planning schemes. Each planning scheme includes the following information. These correspond to key indicators of urban village redevelopment, such as the proportion of area redeveloped, the proportion of commercial housing development, the proportion of rental housing development, the proportion of increased public facilities, and the plot ratio. Since urban village redevelopment does not involve a reverse process, therefore... In addition, external parameters also include the expected growth rate of housing prices. Compared with the average annual inflation rate .

[0084] The plan aims to achieve the city's overall goals of "maintaining growth, addressing weaknesses, and adjusting the economic structure." "Maintaining growth" refers to the economic revenue and expenditure situation before and after the redevelopment, comprised of profits from commercial housing development and rental income from rental housing, with the cumulative total reaching [a certain percentage]. The annual economic return can be calculated using the following formula (1):

[0085] (1)

[0086] Where n represents the index of each plot or area in the urban village.

[0087] ∑n: represents summing over all plots or regions.

[0088] Area n : Represents the area of ​​the nth plot.

[0089] : Represents the completion rate of the transformation of the nth plot after m years.

[0090] : Represents the economic development rate after m years.

[0091] : Represents the plot ratio of the nth plot after m years.

[0092] : Represents the unit price in economic activities, such as real estate prices.

[0093] η1 and η2: These represent parameters related to economic development and inflation rates.

[0094] : Represents the rental rate after m years.

[0095] : represents the product of x=0 to x=30.

[0096] "Addressing shortcomings" refers to the utilization status of public service facilities after renovation. It is calculated using Moran's index, which measures the number of visits to public service facilities between different grids, and can be obtained through the following formula (2):

[0097] (2)

[0098] Where i and j: These are typically the dimensions of the spatial grid, representing the number of grids into which the city is divided. For example, if the city is divided into grids with i rows and j columns, then i×j is the total number of grids.

[0099] These are summation symbols, indicating that the summation is performed over all grid cells.

[0100] This is a weight matrix, where This represents the weight between grid a and grid b. This weight may be based on the distance between the grids, connectivity, or other relevant factors.

[0101] and These represent the point of interest vectors in grids a and b. These vectors may contain information about facilities in the grid, such as facility type, number, or frequency of use.

[0102] This represents the utilization of urban facilities after m years. It may be a vector containing the utilization of facilities in each grid.

[0103] This represents the average vector of the points of interest across all grids.

[0104] "Structural adjustment" refers to the commuting situation within the city after the renovation, which is calculated using the total commuting volume between grids, as shown in the following formula (3):

[0105] (3)

[0106] It was noted that after the redevelopment, the population distribution within the city changed. Facility layout The factors involved in such situations are too complex, therefore It cannot be directly exported from the modification plan.

[0107] Therefore, the core objective of this invention is to train the first generative model. It can derive specific renovation plans based on the existing urban conditions. The city's condition after the new year is then analyzed, and the three main indicators under this scenario are calculated using the formula described above.

[0108]

[0109] Then, by training the second generative model It can be adapted to the conditions of the city after the renovation. This predicts access to urban facilities and internal commuting in this scenario.

[0110]

[0111] Finally, through heuristic algorithm models Adjustment and renovation plan The urban village redevelopment plan aims to maximize overall benefits.

[0112]

[0113] Specifically, see Figure 1 and Figure 2 . Figure 1 This is one of the flowcharts illustrating the urban village redevelopment planning method based on generative models provided by this invention. Figure 2 This is a schematic diagram of the planning framework for the urban village renovation method provided in an embodiment of the present invention.

[0114] The method includes the following:

[0115] Step 101: Based on the determined urban conditions g, urban village conditions u, and redevelopment scheme set P, use the pre-trained first generative model to generate the urban activity status g m years later. m .

[0116] The first generative model is trained based on a set of urban condition training samples, urban village condition samples, and renovation plan sample sets.

[0117] Specifically, data collection and preprocessing include:

[0118] Urban Condition Data: This involves collecting detailed data on the current state of the city, including demographic data, land use types, traffic flow, and economic activity data. This data can be obtained from municipal departments, traffic management bureaus, statistics bureaus, and other sources.

[0119] Data on the conditions of urban villages: Collect detailed data on the conditions of urban villages, including building density, infrastructure status, living conditions, and environmental sanitation. This data can be obtained through on-site surveys, satellite remote sensing, and drone photography.

[0120] Transformation Plan Set P: This set of transformation plans includes planned building projects, infrastructure improvement plans, and estimated investment amounts. These plans can be provided by the city planning department or developed based on expert consultation and public participation.

[0121] The collected data is integrated into a unified data format for input into the generative model. Data preprocessing, including normalization, outlier removal, and missing data imputation, is performed to improve the model's predictive accuracy.

[0122] Load the first pre-trained generative model GenAI α (·), and set the model parameters, including the prediction time range of m years. Input the preprocessed urban conditions, urban village conditions, and redevelopment scheme set into the model. Based on the input data and the learned patterns, the model generates the urban activity status g m years later. m The predicted urban activity status g m It is applied in fields such as urban planning, transportation planning, and infrastructure layout, providing scientific basis for decision-makers.

[0123] Step 101 can provide data-driven future predictions for urban village redevelopment planning, help planners assess the potential impact of different redevelopment schemes, and provide a scientific basis for subsequent planning decisions.

[0124] Step 102: Based on the city's activity status m years later (g) m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m .

[0125] It should be noted that generative models can take many forms, such as neural network models, like Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Autoregressive models. These models are based on neural network architectures and can learn the complex distribution of data and generate new data instances. Probability-based models, such as Hidden Markov Models (HMMs) and Bayesian networks, simulate the distribution of data by defining the probabilistic process of data generation. Optimization-based models, such as some energy-based models (EBMs), learn the distribution of data by optimizing energy functions.

[0126] In this embodiment, the first generative model is a neural network model, and the second generative model is a Markov generative model.

[0127] In this embodiment, the urban activity status g after m years is used as the basis. mThis involves extracting relevant urban infrastructure data, including facility type, location, and capacity. It also includes collecting demographic data, employment information, and lifestyle habits of urban residents; this data will serve as attributes for the agent. Finally, it defines behavioral parameters for the agent in the second generative model, including exploration probability, return probability, and movement cost. These parameters can be determined through statistical analysis of historical data, expert consultation, or pre-experiments.

[0128] Agent locations are initialized randomly or according to specific rules within a city's spatial grid. Each agent is assigned an initial state, including its current grid location and visited city facilities. A second generative model is then used to simulate agent movement within the city, with each agent moving between city grids based on defined parameters. Agent movement may be influenced by factors such as the distribution of city facilities, transportation networks, and individual preferences.

[0129] Based on the agent's movement trajectory and access records, the number of times each city facility was accessed is counted. The utilization rate of city facilities after m years is estimated using the calculation method built into the second generative model. m .

[0130] For the generated I m The results are evaluated to check whether they meet the expected urban development goals and planning requirements. Based on the evaluation results, model parameters are adjusted or the model structure is optimized to improve the accuracy and reliability of the predictions.

[0131] Through step 102, the second generative model It can provide urban planners with in-depth insights into the future utilization of urban facilities, helping them to develop more scientific and reasonable urban village redevelopment plans.

[0132] Step 103: Obtain the economic income and expenditure situation E after m years based on the set of renovation schemes P. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m By using heuristic algorithms for multiple iterations, several adjusted target modification schemes were obtained.

[0133] Initialize the renovation scheme set: Starting from the data obtained in steps 101 and 102, initialize the renovation scheme set P, which includes all possible renovation schemes and their corresponding economic, commuting and facility utilization forecast data.

[0134] Define evaluation indicators: Determine the indicators for evaluating the renovation plan, including but not limited to economic benefits, commuting efficiency, and facility utilization rate. Assign weights to each indicator to reflect its importance in the overall plan.

[0135] In this embodiment, a suitable heuristic algorithm, such as a genetic algorithm, simulated annealing, or particle swarm optimization, can be selected to search for the optimal modification scheme. Algorithm parameters, such as population size, number of iterations, crossover rate, and mutation rate, are set.

[0136] The set of modification schemes P is iteratively optimized using a heuristic algorithm.

[0137] 1) In each generation, evaluate the fitness of each scheme, i.e., its performance on evaluation metrics. Select the best scheme based on fitness as the parent, and generate new offspring schemes through crossover and mutation operations.

[0138] 2) Perform non-dominated ordination on the population for each generation to identify non-dominated modification schemes. For dominated schemes, calculate their crowding to maintain population diversity.

[0139] 3) Select the parent scheme for the next generation based on non-dominated sorting and crowding selection. Combine the selected parent scheme and the generated child schemes to form a new set of modification schemes for the next round of iteration.

[0140] Repeat steps 1) to 3) until the preset number of iterations is met or the set of modification schemes no longer has significant updates.

[0141] Evaluate the optimal solution set for each generation and monitor the convergence performance of the algorithm.

[0142] After the final iteration, the non-dominated solution set within the population is output. This set represents the best-balanced modification plan among multiple objectives. The optimal modification plan is then provided to the decision-maker to support their final planning decision.

[0143] Step 103 provides a systematic framework for optimizing urban village redevelopment planning schemes through heuristic algorithms, ensuring that the schemes achieve the best balance among multiple objectives such as economy, society, and environment.

[0144] Step 104: Analyze the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years to obtain the optimal target renovation scheme.

[0145] In this embodiment, for each adjusted target renovation scheme, the corresponding urban activity status g after all m years is calculated. m Urban commuting situation after m years C m and the utilization of urban facilities after m years I mStatistical processing is performed separately to obtain the total economic revenue and expenditure E, the total urban facility utilization I, and the total urban commuting C within m years. Then, based on the total economic revenue and expenditure E, the total urban facility utilization I, and the total urban commuting C within m years and their respective weights, the optimal target renovation plan is determined.

[0146] A weighted summation method is used to calculate the overall score for each modification scheme. For example, for each scheme P... i Its overall score is S i This can be expressed as formula (4):

[0147] Si=w E ·E i +w I ·I i +w C ·C i (4)

[0148] Among them, w E w I w C These are the weights of economic, facility utilization, and commuting indicators, respectively. i I i C i It is scheme P i Scores on these metrics.

[0149] Step 104 provides a comprehensive analysis and decision-making framework to help decision-makers select the optimal target renovation plan from multiple adjusted renovation options and to support the implementation and subsequent evaluation of the plan.

[0150] The urban village redevelopment planning method based on generative models provided in this invention generates the urban activity status g m years later using a pre-trained first generative model, based on the determined urban status g, urban village status u, and redevelopment scheme set P. m Then, based on the urban activity status g after m years m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m Then, using the economic balance and expenditure situation E after m years m Urban commuting situation after m years C m and the utilization of urban facilities after m years I mThis method utilizes heuristic algorithms for multiple iterations to obtain several adjusted target renovation schemes, and then selects the optimal one. By employing a generative model, this embodiment can quickly simulate and predict the future economic, social, and environmental impacts of different renovation schemes. This allows for the simultaneous consideration of multiple objectives, such as economic benefits, utilization of public service facilities, and urban commuting. Through multi-objective optimization, it finds the optimal balance point, providing decision-makers with more comprehensive and accurate data support, thereby enabling more scientific and rational planning decisions.

[0151] Furthermore, the first generative model and the second generative model involved in this embodiment are illustrated in an illustrative manner.

[0152] The first generative model includes a first network and a second network. The first network is U-Net, and the second network is ResNet.

[0153] Generative models essentially fit a given data sample to a known probability model, thereby drawing a new instance consistent with the given data sample from a known distribution. Diffusion generative models are a branch of generative models. Their basic idea is to gradually add noise to the data through a diffusion process. When there is enough noise, the sample is corrupted into noise (the distribution is known). By fitting this diffusion process through a first network, the noise added at each step can be predicted.

[0154] In the denoising process, starting with random noise, a pre-trained first network is used to predict the noise to be added to the image. By removing the noise, a noise-free image is gradually restored. Simultaneously, to guide the denoising process and correctly generate the urban village condition U and the urban morphology under the given renovation plan P, a corresponding second network is also trained to select the correct direction for noise removal.

[0155] See Figure 3 and Figure 4 The training method for the first network includes:

[0156] 301. Randomly select a city condition training sample from the training sample dataset. .

[0157] A sample is randomly drawn from a pre-collected training dataset of urban conditions. Ensure that sample g0 contains sufficient historical information to represent the state of the city at a specific point in time.

[0158] 302. Determine the random sampling time t, and from a constant sequence Extract the corresponding constant from .

[0159] Determine a random sampling time point, which can be selected based on a strategy (such as uniform distribution, specific event triggering, etc.). Extract a constant corresponding to the sampling time from a predefined sequence of constants; this constant may represent the influence weight of time or other time-related coefficients.

[0160] A random time interval m is selected, which defines the time range from the current time point to the beginning of the sample dataset.

[0161] 303. From a training sample related to urban conditions From a standard Gaussian distribution of the same size, a random sample of noise is randomly drawn. and random sample noise and constants Add to urban condition training samples In the process, we obtain training samples of urban conditions after adding noise. .

[0162] Among them, the noisy urban condition training samples Calculated using the following formula (5):

[0163] (5)

[0164] 304. The noisy urban condition training samples With the time of random sampling The input is fed into the first network to obtain the predicted noise. .

[0165] 305. Calculate the loss value based on the predicted noise and random sample noise, and update the parameters in the first network using the predicted loss value through stochastic gradient descent until the parameters in the first network converge.

[0166] Wherein, the loss function is

[0167] Through steps 301-305, the first network of the first generative model can learn to predict noise from noisy urban conditions and gradually reconstruct the original urban activity situation. This training method helps improve the model's performance in the denoising process, thereby generating more accurate predictions of future urban conditions and providing more reliable data support for urban village redevelopment planning.

[0168] See Figure 4 and Figure 5 The training method for the second network includes:

[0169] 501. Randomly select a city condition training sample from the training sample dataset. .

[0170] 502. Training samples of randomly selected city conditions The corresponding time point is taken as the starting point, and the process continues forward until the starting time point of the sample dataset is taken as the ending point. A time interval m is randomly selected.

[0171] 503. Based on the sampling time interval m, obtain the known urban village status sample U0 and the corresponding renovation plan sample set P. θ .

[0172] 504. Randomly selected urban condition training samples The predicted modification time interval is obtained by inputting the data into the second network. Urban Village Situation and renovation plan P θ .

[0173] 505. Calculate the predicted renovation time interval. The time interval between sampling and the predicted conditions of urban villages Compared with known samples of urban village conditions and predicted redevelopment plans P θ The second loss value is obtained by comparing the difference with the known modification scheme samples.

[0174] The loss function is: .

[0175] 506. Using stochastic gradient descent, update the parameters of the second network with the second loss value until the parameters of the second network converge.

[0176] The second network of the first generative model can learn to predict future redevelopment effects from the current state of the city and optimize model parameters based on historical data, thereby improving the accuracy of the model's predictions of future urban activity. This training method helps generate more accurate predictions of future urban conditions, providing more reliable data support for urban village redevelopment planning.

[0177] See Figure 6 Step 101 specifically includes:

[0178] 601. Based on the determined city condition g, extract the noise quantity from the standard Gaussian distribution, and then add it to the determined city condition g to obtain the current random city condition. .

[0179] 602. Random city conditions Compared with the predefined maximum number of denoising steps The input is fed into a pre-trained first network for processing to obtain the predicted noise. .

[0180] 603. Random city conditions The conditions of the urban village (u) and the set of renovation plans (P) are input into a pre-trained second network to calculate the gradient of the loss function. .

[0181] 604. Update the noise based on the gradient value. The current random city status is updated based on the updated noise, resulting in the updated random city status.

[0182] The updated noise is calculated using the following formula (6):

[0183] (6)

[0184] The updated random city conditions are calculated using the following formula (7):

[0185] (7)

[0186] 605. Iteratively execute the step of inputting the updated random city conditions into the first network and the second network respectively to obtain the next updated random city conditions, until the maximum number of denoising steps T is 0, and obtain the city activity status after m years.

[0187] Following the steps outlined above, we only need to specify the maximum number of noise reduction steps. and constant sequence That is, the first generative model can be obtained. There are two related first networks Second network Together, they can be adapted to the existing urban conditions. Urban Village Situation and the given renovation plan ,generate City conditions after the New Year .

[0188] In addition, for the second generative model It can adapt to the urban conditions after the renovation. m Predicting urban facility access in this scenario and internal commuting situation For its general framework, see [link / reference]. Figure 7 .

[0189]

[0190] The core of this problem lies in the fact that the factors influencing commuting in cities are highly complex, making it difficult to predict urban commuting conditions as a whole. In this patent, we use an exploration-memory-movement behavior-based method to generate the behavior of each micro-entity in the city and sum them up to obtain the facility access and commuting conditions of the entire city.

[0191] Existing research indicates that human movement in cities follows a power-law distribution, meaning long-distance travel is infrequent and short-distance travel is frequent, with a small number of destinations accounting for the majority of trips, while most destinations see only a small number of trips. This phenomenon is attributed to the fact that human activity in cities is both memory-based (influenced by previous activities) and social (influenced by the activities of other humans). Therefore, we employ a two-stage approach, simultaneously considering the return of humans to previously visited locations and their interactions with other humans. The basic process of generation is as follows Figure 8 As shown:

[0192] 801. Define multiple agents in a city. For any agent in the city... It records the set of all the locations it has visited.

[0193] 802. For any time Random numbers are drawn from a random distribution. .

[0194] 803. Calculate the exploration probability of agent i. .

[0195] in, and These are two pre-determined hyperparameters. It represents the number of grid points that agent i has visited.

[0196] 804. If p≥ The control agent performs a return operation, randomly selecting a location from the set of locations it has visited and updating its location.

[0197] The probability of visiting each location is directly proportional to the frequency with which it has been visited previously: .

[0198] 805. If The control agent i performs an exploration operation, randomly selecting a location from the set of other locations in the city and updating its location; wherein, the probability of visiting each location is directly proportional to the number of people in that location, and the probability of visiting each location is inversely proportional to the distance between two locations.

[0199] Specifically, see the following formula (8):

[0200] (8)

[0201] in, These are the starting location and a random location, respectively. The distance between two locations; These are the core parameters.

[0202] 806. Perform the above steps on all agents in the city until time. Reaching the specified time step .

[0203] 807. Count the number of visits to the city facilities of interest, and calculate the utilization of the city facilities after m years.

[0204] 808. Count the number of passages between grids within the city, and calculate the commuting situation within the city after m years. .

[0205] Note the core parameters in this process. All of these are unknown, and there may be individual differences. This patent constructs a Markov generative model, which uses the maximum likelihood method to solve for the parameters that make the observed sample most likely to occur during a portion of human movement, thereby generating possible samples.

[0206] Assuming the core parameters of the model All are from standard Gaussian distribution .

[0207] Based on existing sequences of human behavior patterns The frequency of exploration behavior, return behavior, and visited locations of all samples were statistically analyzed, and the core parameters of each sample were calculated. ;

[0208] Based on the statistically obtained parameters, calculate the log-likelihood function that makes this series of samples appear.

[0209] Using the obtained probability distribution Based on the given city conditions Extract and generate a certain proportion of proxies for each grid, and execute steps 801-808 above.

[0210] Optionally, see Figure 9 and Figure 10 . Figure 9 This is a schematic diagram of the framework of the heuristic algorithm H. Figure 10 This is a flowchart illustrating step 103. Specifically, step 103 includes:

[0211] 1001. Based on the economic income and expenditure situation after m years, E mUrban commuting situation after m years C m and the utilization of urban facilities after m years I m The dominance relationships between objective functions are calculated to obtain non-dominated and dominated modification schemes.

[0212] Before executing step 1001, the relevant parameters of the renovation plan are... The upper and lower boundaries are all set by experts. and Set the population size. iteration rounds Crossover probability Probability of mutation Random initialization A set of different renovation plans yields the solution set. .

[0213] 1002. For the dominated modification schemes, according to the level of domination, the crowding degree is sorted within each level based on the distance between the dominated modification schemes.

[0214] 1003. Based on the degree of domination and crowding of the dominated transformation schemes, select the... The optimal modification scheme is used as the parent solution for the next generation.

[0215] 1004. Calculate the crossover probability between parent trees. Mutation probability The same result can be obtained by exchanging or updating parts of the sequence of the modification scheme. The modification plan for each offspring is used as the solution for the next generation of offspring.

[0216] The offspring solutions generated through crossover and mutation operations are not guaranteed to be "optimal." They are new candidate solutions that need to be evaluated along with other individuals in the next generation's selection operation. Only those offspring solutions that perform well in the evaluation are selected as the parents for the next generation. This process helps the algorithm explore the search space and gradually approach the true Pareto optimal solution set.

[0217] 1005. Using the parent solution and the child solution of the next generation as the updated modification scheme, process them through the first generative model and the second generative model to obtain the updated economic income and expenditure situation E after m years. m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m It continues to iterate until the required number of iterations is met or the set of modification schemes P is no longer updated.

[0218] 1006. Output the results of the last iteration, obtaining multiple adjusted non-dominated target modification schemes.

[0219] In each generation of the genetic algorithm, a new generation of population is generated through selection, crossover, and mutation operations, consisting of k parent solutions and k offspring solutions. These solutions represent the best modification plans in the current population, achieving a balance among multiple objectives.

[0220] The above heuristic algorithm model Search within the given solution space for the set of non-dominated solutions that represent the optimal transformation schemes for that year. Typically, a single solution is not obtained, but rather a set of solutions, each with its own advantages and disadvantages in different directions. These solutions are all feasible and excellent modification schemes, with varying preferences for different modification goals, allowing decision-makers to compare and select the best option during the actual process.

[0221] Taking City A as an example, to generate a suitable planning scheme for the urban village renovation in City A, the following methods can be used:

[0222] (1) The city of A is divided into spatial grids with 1200 meters as one grid and distances of 99 and 125 in the latitude and longitude directions, respectively. After removing non-urban areas, a total of 2267 grids are obtained.

[0223] (2) Extract the urban village plots in City A and obtain a total of 1,481 urban villages to be renovated and related information.

[0224] (3) Use historical data on urban village renovation in areas A, B, and C to train generative models. The time is set to be sufficiently long. , ,in .in, It is a time-dependent parameter that controls the noise level at each step in the denoising process; is another time-dependent parameter that determines the proportion of noise removed in each step; T represents the maximum number of steps in the denoising process, which is a hyperparameter that can be adjusted according to the complexity of the model and the characteristics of the training data.

[0225] (4) Use the behavioral activity sequences of people in City A to train a generative model. To accelerate computation, the ratio of model-generated agents to real population is 1:1000.

[0226] (5) Set the parameters required for operation, including the solution space setting as follows: 0.6 <r a <1, 0.5 <r e <0.8, 0.15 <rr <0.25, 1.3 < r poi <1.5, 1.8 < FAR < 2.6; The total renovation years m = 10 for the external parameter setting, η1 = 1.05, η2 = 1.02; The internal parameters of the heuristic model k = 50, g = 10, p c = 0.8, p m = 0.5.

[0227] 0.6 < r a <1: represents the renovation completion rate r a range. This ratio may reflect the proportion of urban villages that have completed renovation within a given time period.

[0228] 0.5 < r e <0.8: represents the economic development rate r e range. This ratio may be related to the urban economic growth or the development speed of a specific area.

[0229] 0.15 < r r <0.25: represents the rental rate r r range. This ratio may reflect the change in the rental level of the renovated area.

[0230] 1.3 < r poi <1.5: represents the utilization of urban facilities r poi range. This ratio may be related to the usage frequency or efficiency of urban facilities.

[0231] 1.8 < FAR < 2.6: represents the range of the floor area ratio FAR. The floor area ratio is a measure of building density and reflects the total amount of buildings permitted on a given land area.

[0232] (6) According to the method described above, using the heuristic model , call the trained first generative model and the second generative model , to obtain 4 non-dominated renovation sets

[0233] In this embodiment, there are 3 evaluation dimensions (development benefit, facility configuration, traffic evaluation), so select the best-performing renovation plan set in each dimension, and add a renovation plan set that balances between various goals, so there are 4 non-dominated renovation plan sets.

[0234] (7) Analyze the respective situations of the 4 non-dominated renovation plan sets, and conduct corresponding naming and analysis based on the comparison between their scores.

[0235] By selecting the optimal renovation plan, limited financial and construction resources can be allocated more effectively, improving resource utilization efficiency. Choosing a plan that considers facility configuration and commuting conditions can enhance the overall service functions of the city and improve the quality of life for residents.

[0236] The following describes the urban village renovation planning system based on generative models provided in the embodiments of the present invention. The urban village renovation planning system based on generative models described below and the urban village renovation planning method based on generative models described above can be referred to and correspond to each other.

[0237] This invention provides an urban village redevelopment planning system based on a generative model. (See also...) Figure 11 ,include:

[0238] The first parameter generation module 111 is used to generate the urban activity status g m years later, based on the determined urban status g, urban village status u, and redevelopment scheme set P, using a pre-trained first generative model. m The first generative model is trained based on urban condition training samples, urban village condition samples, and renovation scheme sample sets.

[0239] The second parameter generation module 112 is used to generate the city activity status g m years later. m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m ;

[0240] Iterative calculation module 113 is used to obtain the economic income and expenditure situation E after m years based on the set of transformation schemes P. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Multiple adjustments to the target modification scheme were obtained by using heuristic algorithms for multiple iterations.

[0241] The target renovation scheme generation module 114 is used to analyze the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years to obtain the optimal target renovation scheme.

[0242] The system in this embodiment uses a generative model to quickly simulate and predict the future economic, social, and environmental impacts of different renovation schemes. This allows for the simultaneous consideration of multiple objectives, such as economic benefits, utilization of public service facilities, and urban commuting. By optimizing these multiple objectives, the system finds the optimal balance point, providing decision-makers with more comprehensive and accurate data support, thereby enabling them to make more scientific and rational planning decisions.

[0243] Figure 12 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 12 As shown, the electronic device may include: a processor 1210, a communications interface 1220, a memory 1230, and a communication bus 1240. The processor 1210, communications interface 1220, and memory 1230 communicate with each other via the communication bus 1240. The processor 1210 can call logical instructions in the memory 1230 to execute a generative model-based urban village redevelopment planning method. This method includes: based on the determined urban conditions g, urban village conditions u, and redevelopment scheme set P, using a pre-trained first generative model to generate the urban activity status g m years later. m The first generative model is trained using training samples of urban conditions, samples of urban village conditions, and a sample set of redevelopment plans; the model is then trained based on urban activity data (g) after m years. m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m Based on the set of renovation plans P, the economic income and expenditure situation E after m years is obtained. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Multiple adjustments to the target renovation scheme are obtained through iterative processing using a heuristic algorithm. Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years, the optimal target renovation scheme is obtained.

[0244] Furthermore, the logical instructions in the aforementioned memory 1230 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0245] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the urban village redevelopment planning method based on generative models provided by the above methods. The method includes: generating the urban activity status g m years later using a pre-trained first generative model based on the determined urban status g, urban village status u, and redevelopment scheme set P. m The first generative model is trained using training samples of urban conditions, samples of urban village conditions, and a sample set of redevelopment plans; the model is then trained based on urban activity data (g) after m years. m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m Based on the set of renovation plans P, the economic income and expenditure situation E after m years is obtained. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I mMultiple adjustments to the target renovation scheme are obtained through iterative processing using a heuristic algorithm. Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years, the optimal target renovation scheme is obtained.

[0246] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the urban village redevelopment planning method based on generative models provided by the above methods. This method includes: generating the urban activity status g m years later using a pre-trained first generative model, based on the determined urban status g, urban village status u, and redevelopment scheme set P. m The first generative model is trained using training samples of urban conditions, samples of urban village conditions, and a sample set of redevelopment plans; the model is then trained based on urban activity data (g) after m years. m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m The second generative model assumes multiple agents moving within the city and uses the maximum likelihood method to solve for the movement parameters of these agents. This allows it to predict the intra-city commuting situation C m years later, based on the modified city conditions. m and the utilization of urban facilities after m years I m Based on the set of renovation plans P, the economic income and expenditure situation E after m years is obtained. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Multiple adjustments to the target renovation scheme are obtained through iterative processing using a heuristic algorithm. Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years, the optimal target renovation scheme is obtained.

[0247] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0248] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0249] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A planning method for urban village redevelopment based on a generative model, characterized in that, include: According to the determined urban condition g, the village-in-city condition u and the reconstruction scheme set P, an urban activity condition g after m years is generated by using a pre-trained first generative model m ; wherein the first generative model is trained according to the urban condition training sample, the village-in-city condition sample and the reconstruction scheme sample set; According to the city activity situation g after m years m , using a preset second generative model, calculate the city internal commuting situation C after m years m And the city facility utilization situation I after m years m ; wherein, the second generative model is solved by assuming that a plurality of agents move in the city, and the moving parameters of the plurality of agents in the city are solved by maximum likelihood method, so as to predict the city internal commuting situation C after m years m And the city facility utilization situation I after m years m under the situation according to the modified city situation; According to the reconstruction scheme set P, economic revenue and expenditure E after m years is obtained m , according to the economic revenue and expenditure E after m years m , the urban internal commuting C after m years m and the urban facility utilization I after m years m , a plurality of adjusted target reconstruction schemes are obtained by using a heuristic algorithm for multiple iterations; Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years, the optimal target renovation scheme is obtained. The first generative model includes a first network and a second network; According to the determined urban condition g, the urban-village condition u, and the reconstruction scheme set P, an urban activity condition g after m years is generated by using a pre-trained first generative model m , and specifically includes: Given the established city condition g, noise is extracted from a standard Gaussian distribution and then added to the established city condition g to obtain the current random city condition. Random city conditions The predefined maximum number of denoising steps T is input into a pre-trained first network for processing to obtain the predicted noise ∈ θ ; Random city conditions The conditions of the urban village (u) and the set of renovation plans (P) are input into a pre-trained second network to calculate the gradient of the loss function. updating the noise ε according to the gradient value θ and updating the current random city condition according to the updated noise to obtain an updated random city condition; The process iteratively executes the step of inputting the updated random city conditions into the first network and the second network respectively to obtain the next updated random city conditions, until the maximum number of denoising steps T is 0, and obtains the city activity status after m years; Based on the urban activity situation m years later g m Using a pre-defined second generative model, calculate the intra-city commuting situation C after m years. m and the utilization of urban facilities after m years I m Specifically, it includes: Define multiple proxies in a city. For any proxies i in the city, record the set of all locations they have visited. For any time t, take a random number p from a random distribution; Exploration probability of agent i, δS -γ where δ and γ are two predetermined hyperparameters, and S is the number of grid points that agent i has visited; If p≥δS -γ , the control agent performs a return operation, randomly selects a location from the set of locations it has visited, and updates its position; wherein the probability of selecting each location is proportional to the frequency with which it has been visited previously. if p < δS -γ , the control agent i performs an exploration operation, randomly selects a location from the other location set in the city, and updates its position; wherein the access probability of each location is proportional to the number of people in the location, and the access probability of each location is inversely proportional to the distance between two locations; Perform the above steps on all agents in the city until time t reaches the specified time step T; counting the number of visits to the city facilities of interest, calculating the city facility utilization I after m years m ; counting the number of passages between the grids inside the city, calculating the intra-city commute C after m years m .

2. The urban village redevelopment planning method based on generative models according to claim 1, characterized in that, The training method for the first network includes: Randomly select a city condition training sample from the training sample dataset; determining a random extraction time t and taking the corresponding constant a from a constant sequence {a T} of constants t ; A random sample noise is randomly drawn from a standard Gaussian distribution of the same size as the city condition training sample, and the random sample noise and the constant are added to the city condition training sample to obtain the noisy city condition training sample. The noisy urban condition training samples and the randomly selected time t are input into the first network to obtain the predicted noise; The loss value is calculated based on the predicted noise and random sample noise. The parameters in the first network are updated using the predicted loss value through stochastic gradient descent until the parameters in the first network converge.

3. The urban village redevelopment planning method based on generative models according to claim 1, characterized in that, The training methods for the second network include: Randomly select a city condition training sample from the training sample dataset. Training samples of randomly selected urban conditions The corresponding time point is taken as the starting point, and the process continues forward until the time point at the beginning of the sample dataset is taken as the ending point. A time interval m is randomly selected. Based on the sampling time interval m, obtain a sample of the known conditions of urban villages. and the corresponding set of renovation scheme samples Randomly selected urban condition training samples Inputting the data into the second network yields the predicted modification time interval m. θ Urban Village Situation and renovation plan Calculate the predicted modification time interval m θ The time interval between sampling and the predicted conditions of urban villages Compared with known samples of urban village conditions and predicted redevelopment plans The difference between the sample and the known modification schemes is used to obtain the second loss value; The parameters of the second network are updated using the second loss value through stochastic gradient descent until the parameters of the second network converge.

4. The urban village redevelopment planning method based on generative models according to claim 1, characterized in that, According to the economic income and expenditure E after m years m , the urban internal commuting C after m years m , and the urban facility utilization I after m years m , a plurality of adjusted target reconstruction schemes are obtained through a plurality of iterations by using a heuristic algorithm, and the plurality of adjusted target reconstruction schemes specifically include: According to the economic income and expenditure E after m years m , the intra-city commuting C after m years m , and the utilization of urban facilities I after m years m , a domination relationship between the target functions is calculated to obtain non-dominated reconstruction schemes and dominated reconstruction schemes; For the dominated modification schemes, the crowding is sorted according to the distance between the dominated modification schemes within each level, based on the degree of domination. Based on the degree of domination and crowding of the dominated transformation schemes, select k optimal transformation schemes as the parent solutions for the next generation; Calculating the probability of crossing p between the parents c with the probability of variation p m , exchanging or updating the partial sequence of the retrofitting solution, obtaining again k offspring of the retrofitting solution as offspring solutions of the next generation; The parent solution of the next generation and the offspring solution of the next generation are taken as the updated reconstruction scheme, and are processed through the first generative model and the second generative model to obtain the updated economic income and expenditure situation E after m years m , the urban internal commuting situation C after m years m , and the urban facility utilization situation I after m years m The iteration is continuously performed until the iteration round is satisfied or the reconstruction scheme set P is no longer updated. Output the results of the last iteration, resulting in multiple adjusted non-dominated target modification schemes.

5. The urban village redevelopment planning method based on generative models according to claim 1, characterized in that, Based on the analysis of the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban intra-city commuting C within m years, the optimal target renovation scheme is obtained, which specifically includes: For each adjusted target transformation scheme, the corresponding total urban activity status g m , the total urban internal commuting status C m and the total urban facility utilization status I m after m years are respectively statistically processed, and the total economic income and expenditure E, the total urban facility utilization status I and the total urban internal commuting status C within m years are respectively obtained. The optimal target renovation plan is determined by calculating the total economic revenue and expenditure (E) over m years, the total urban infrastructure utilization (I) over m years, and the total urban commuting situation (C) over m years, along with their respective weights.

6. A planning system for urban village redevelopment based on a generative model, characterized in that, include: The first parameter generation module is configured to generate the urban activity condition g after m years by using a pre-trained first generative model according to the determined urban condition g, the village-in-city condition u, and the reconstruction scheme set P. m ; wherein the first generative model is trained according to an urban condition training sample, a village-in-city condition sample, and a reconstruction scheme sample set. a second parameter generation module configured to calculate, according to the city activity situation g after m years m , the city internal commuting situation C after m years and the city facility utilization situation I after m years by using a preset second generative model m m ; wherein the second generative model is configured to solve the moving parameters of a plurality of agents in the city by assuming that the plurality of agents move in the city by using a maximum likelihood method, so as to predict the city internal commuting situation C after m years and the city facility utilization situation I after m years according to the city situation after the reconstruction m m ;​​ The iterative calculation module is used to obtain the economic income and expenditure situation E after m years based on the set of transformation schemes P. m Based on the economic balance and expenditure situation after m years, E m Urban commuting situation after m years C m and the utilization of urban facilities after m years I m Multiple adjustments to the target modification scheme were obtained by using heuristic algorithms for multiple iterations. The target renovation scheme generation module is used to analyze the multiple adjusted target renovation schemes and their corresponding total economic revenue and expenditure E, total urban facility utilization I, and total urban commuting C within m years to obtain the optimal target renovation scheme. The first generative model includes a first network and a second network; The first parameter generation module is specifically used for: Given the established city condition g, noise is extracted from a standard Gaussian distribution and then added to the established city condition g to obtain the current random city condition. Random city conditions The predefined maximum number of denoising steps T is input into a pre-trained first network for processing to obtain the predicted noise ∈ θ ; Random city conditions The conditions of the urban village (u) and the set of renovation plans (P) are input into a pre-trained second network to calculate the gradient of the loss function. Update the noise ∈ based on the gradient value. θ The current random city status is updated based on the updated noise to obtain the updated random city status. The process iteratively executes the step of inputting the updated random city conditions into the first network and the second network respectively to obtain the next updated random city conditions, until the maximum number of denoising steps T is 0, and obtains the city activity status after m years; The second parameter generation module is specifically used for: Define multiple proxies in a city. For any proxies i in the city, record the set of all locations they have visited. For any time t, take a random number p from a random distribution; Calculate the exploration probability δS of agent i. -γ , where δ and γ are two pre-determined hyperparameters, and S is the number of grid points that agent i has visited; If p≥δS -γ The control agent performs a return operation, randomly selecting a location from the set of locations it has visited and updating its location; where the probability of visiting each location is proportional to the frequency of its previous visits; If p < δS -γ The control agent i performs an exploration operation, randomly selecting a location from the set of other locations in the city and updating its location; wherein, the probability of visiting each location is directly proportional to the number of people in that location, and the probability of visiting each location is inversely proportional to the distance between two locations; Perform the above steps on all agents in the city until time t reaches the specified time step T; Count the number of visits to city facilities of interest, and calculate the utilization rate of city facilities after m years. m ; Count the number of passages between grids within the city, and calculate the city's commuting situation C after m years. m .

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the urban village redevelopment planning method based on a generative model as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the urban village redevelopment planning method based on a generative model as described in any one of claims 1 to 5.

Images